International Journal of Game Theory

Necessary conditions are established for a point contained in the interior or boundary of a convex feasible subset of Euclidean space to be quasi-undominated in an anonymous simple game. Most of the conditions are behaviorally intuitive and imply pariwise symmetries among utility gradients.

We consider the “and” communication device that receives inputs from two players and outputs the public signal yes if both messages are yes, and outputs no otherwise. We prove that no correlation can securely be implemented using this device, even when infinitely many stages of communication are allowed.

In Borm, Keiding, McLean, Oortwijn and Tijs (1992) the compromise value is introduced as a solution concept on the class of compromise admissible NTU-games. Two characterizations of the compromise value are provided on subclasses of NTU-games. This note shows that in one of these characterizations the axioms are dependent. As a result of this observation a new characterization of the compromise value is provided. Moreover, it turns out that with a small weakening of the symmetry property the axioms in the original characterization become independent. Further, it is shown that these characterizations can be extended to a larger class of NTU-games. Finally, all monotonic, Pareto optimal, and covariant values on this class of NTU-games are described.

We expand Crawford and Sobel’s (Econometrica 50(6):1431–1451, 1982) model of information transmission to allow for the costly provision of ‘hard evidence’ in addition to conventional cheap talk. Under mild assumptions we prove that equilibria have an interval-partition structure, where types of the Sender belonging to the same interval either all induce the same action through cheap talk or reveal their types through hard evidence. We also show that the availability of costly hard signals may reverse one of the important implications of the classical cheap talk model, namely, that diverging preferences always lead to less communication.

The core of ann-person game is the set of feasible outcomes that cannot be improved upon by any coalition of players. A convex game is defined as one that is based on a convex set function. In this paper it is shown that the core of a convex game is not empty and that it has an especially regular structure. It is further shown that certain other cooperative solution concepts are related in a simple way to the core: The value of a convex game is the center of gravity of the extreme points of the core, and the von Neumann-Morgenstern stable set solution of a convex game is unique and coincides with the core.

We state an integer linear programming formulation for the unique characterization of complete simple games, i.e. a special subclass of monotone Boolean functions. In order to apply the parametric Barvinok algorithm to obtain enumeration formulas for these discrete objects we provide a tailored decomposition of the integer programming formulation into a finite list of suitably chosen sub-cases. As for the original enumeration problem of Dedekind on Boolean functions we have to introduce some parameters to be able to derive exact formulas for small parameters. Recently, Freixas et al. have proven an enumeration formula for complete simple games with two types of voters. We will provide a shorter proof and a new enumeration formula for complete simple games with two minimal winning vectors.

We study Nash implementation by natural price–quantity mechanisms in pure exchange economies when agents have intrinsic preferences for responsibility. An agent has an intrinsic preference for responsibility if she cares about truth-telling that is in line with the goal of the mechanism designer besides her material well-being. A semi-responsible agent is an agent who, given what her opponents do, acts in an irresponsible manner when a responsible behavior poses obstacles to her material well-being. The class of efficient allocation rules that are Nash implementable is identified provided that there is at least one agent who is semi-responsible. The Walrasian rule is shown to belong to that class.

An information structure in a non-cooperative game determines the signal that each player observes as a function of the strategy profile. Such information structure is called non-manipulable if no player can gain new information by changing his strategy. A Conjectural Equilibrium (CE) (Battigalli in Unpublished undergraduate dissertation, 1987; Battigalli and Guaitoli 1988; Decisions, games and markets, 1997) with respect to a given information structure is a strategy profile in which each player plays a best response to his conjecture about his opponents’ play and his conjecture is not contradicted by the signal he observes. We provide a sufficient condition for the existence of pure CE in games with a non-manipulable information structure. We then apply this condition to prove existence of pure CE in two classes of games when the information that players have is the distribution of strategies in the population.